Bug Fixes for example notebooks

[pw0908]

It's been a while since we've checked the example notebooks to see if they still run and, in the process of checking them, I found a few bugs:

• cubic_eos.ipynb: Any VTPR examples no longer run with bubble_pressure just giving NaN.

julia> model = VTPR(["carbon monoxide","carbon dioxide"])
PR{BasicIdeal, TwuAlpha, RackettTranslation, VTPRRule{VTPRUNIFAC{BasicIdeal}}} with 2 components:
 "carbon monoxide"
 "carbon dioxide"
Contains parameters: a, b, Tc, Pc, Mw

julia> bubble_pressure(model,218.15,[1e-5,1-1e-5])
(NaN, NaN, NaN, [NaN, NaN])

• GE_mixing_SAFT.ipynb: bubble_temperature no longer takes T0 as an initial guess for activity coefficient models
• user_defined_eos.ipynb: We've updated SAFT-VR Mie and that needs to be reflected here.
• multi-component_vle_vlle_lle_crit.ipynb: There's unfortunately a few bugs here:
  • For the 12-component pT-isopleth, SAFT-gamma Mie struggles to split the model:

julia> model4 = SAFTgammaMie(["methane","nitrogen","carbon dioxide","ethane","propane","butane","isobutane",
                          "pentane","isopentane","hexane","heptane","octane"]);
julia> x =[0.833482,0.056510,0.002840,0.075260,0.020090,0.005200,0.003050,0.001440,0.001200,0.000680,0.000138,0.000110];

julia> bubble_pressure(model4,150.,x)
┌ Error: SAFTVRMie{BasicIdeal} cannot be splitted

• For the T-xxy diagram of methanol+cyclohexane, SRK predicts the wrong vle and lle.

[longemen3000]

the 12 component split was a bug on group_sum. looking on the others